Investigating the Interplay of Self-Construal and Independent Vs. Interdependent Self-Affirmation

Abstract

This study examines how the relative effects of independent and interdependent self-affirmation might be influenced by individuals’ self-construal. A controlled experiment involving 186 Chinese adult smokers revealed a significant interactive effect of self-affirmation type and self-construal on message derogation and posttest attitudes toward smoking. For individuals with a dominant independent self-construal, independent self-affirmation (i.e., affirming the most important value to the self) led to less message derogation and more negative attitudes toward smoking than interdependent self-affirmation (i.e., affirming the most important value to the self and the family). In contrast, for individuals with a dominant interdependent self-construal, interdependent self-affirmation resulted in less message derogation and more negative attitudes toward smoking than independent self-affirmation. In addition, the interaction between self-affirmation type and self-construal had an indirect effect on perceived susceptibility to smoking-related diseases through message derogation. Theoretical and practical implications are discussed.

Notes

¹ Private self refers to “cognitions that involve traits, states, and behaviors of the person.” Public self describes “cognitions concerning the generalized other’s view of the self.” Collective self deals with “cognitions concerning a view of the self that is found in some collective” (Triandis, 1989, p. 507).

² We first identified study subjects that had desired characteristics (i.e., adult smokers) and recruited them to our study. We then asked these subjects to refer other similar subjects to help us recruit more participants (Atkinson & Flint, 2001).

³ We conducted two analyses of variance (ANOVA) to determine if experiment condition affects participants’ self-construal. Experiment condition had no significant effect on the independent self-construal score, F (1, 184) = .338, p = .562, η² = .002, but had a significant effect on the interdependent self-construal score, F (1, 184) = 4.894, p = .028, η² = .026. Surprisingly, the interdependent self-construal score was higher in the independent self-affirmation group (M = 5.24; SE = .12) than the interdependent self-affirmation group (M = 4.88; SE = .11). However, since our focus was not on the relative effects of different types of self-affirmation (i.e., is independent self-affirmation more or less effective than interdependent self-affirmation in reducing defensive processing), but rather whether and how self-construal moderates such effects, the surprising contamination does not substantively change the interpretation of our key results.

⁴ We also ran a simple slope test to probe these interactions. The effect was estimated at three values of self-construal: one standard deviation below the mean of self-construal (= −.3792), the mean (= .8548), and one standard deviation above the mean of self-construal (= 2.0888). A higher score indicates a dominant interdependent self-construal and a lower score indicates a dominant independent self-construal. For message derogation, we found that for participants with a score that was one standard deviation below the mean of self-construal, independent self-affirmation led to less message derogation, b = −.68, SE = .30, p = .03. No significant effect was found at the mean of self-construal and one standard deviation above the mean of self-construal. For posttest attitudes, we found that for participants with a score that was one standard deviation above the mean of self-construal, interdependent self-affirmation led to more negative attitudes toward smoking, b = .46, SE = .19, p = .02. No significant effect was found at the mean of self-construal and one standard deviation below the mean of self-construal.